Software Handling of Hardware Error Handling in Hypervisor-Based Systems

1. A method, comprising: receiving, from a bus of a hypervisor-based system, an error interrupt; identifying at least one affected guest partitions from a plurality of guest partitions executing on the hypervisor-based system; and broadcasting the error interrupt to the affected guest partitions; wherein when the error interrupt corresponds to a correctable error, a service partition processes the error interrupt locally, and when the error interrupt corresponds to an uncorrectable error at a logical CPU, the method further comprises: setting an uncorrectable error bit for each virtual CPU of the logical CPU; determining whether the error interrupt affects any of the virtual CPUs of the logical CPU; and signaling the guest partition corresponding to the one or more of the virtual CPUs of the existence of the error interrupt to enable the guest partition to take corrective action when the error interrupt affects one or more of the virtual CPUs of the logical CPU.

2. The method of claim 1 , in which the step of identifying the affected guest partitions comprises determining owners of a device in the hypervisor system that generated the error interrupt.

3. The method of claim 1 , in which the bus is a peripheral component interconnect express (PCI-e) bus; and the error interrupt comprises a corrupt data error interrupt.

4. The method of claim 1 , wherein the service partition processing the error interrupt comprises: taking corrective action without notifying the guest partitions and logging the error.

5. The method of claim 1 , further comprising preventing reception of further error interrupts from the bus after receiving the error interrupt.

6. The method of claim 1 , further comprising performing a corrective action after broadcasting the error interrupt and the performing the corrective action comprises shutting down the affected guest partitions.

7. A computer program product, comprising: a non-transitory computer readable medium comprising: code to receive, from a bus of a hypervisor-based system, an error interrupt; code to identify at least one affected guest partitions from a plurality of guest partitions executing on the hypervisor-based system; code to broadcast the error interrupt to the affected guest partitions; code to process the error interrupt locally by a service partition when the error interrupt corresponds to a correctable error; and when the error interrupt corresponds to an uncorrectable error at a to CPU, the computer readable medium further comprises: code to set an uncorrectable error bit for each virtual CPU of the logical CPU; code to determine whether the error interrupt affects any of the virtual CPUs of the logical CPU; and code to signal the guest partition corresponding to the one or more of the virtual CPUs of the existence of the error interrupt to enable the guest partition to take corrective action when the error interrupt affects one or ore of the virtual CPUs of the logical CPU.

8. The computer program product of claim 7 , in which the medium further comprises code to determine owners of a device in they hypervisor-based system that generated the error interrupt.

9. The computer program product of claim 7 , in which the bus is a peripheral component interconnect express (PCI-e) bus; and the medium further comprises code to prevent reception of further error interrupts from the bus after receiving the error interrupt.

10. The computer program product of claim 7 , in which wherein the service partition processing the error interrupt comprises: taking corrective action without notifying the guest partitions and logging the error.

11. The computer program product of claim 10 , in which the medium further comprises code to shut down the affected guest partitions.

12. An apparatus, comprising: a bus; and a processor coupled to the bus, in which the processor is configured: to receive from the bus an error interrupt; to identify at least one affected guest partitions from a plurality of guest partitions executing on the processor; to broadcast the error interrupt to the affected guest partitions; and to process the error interrupt locally by a service partition when the error interrupt corresponds to a correctable error; wherein the service partition processing the error interrupt comprises: taking corrective action without notifying the guest partitions and logging the error; and when the error interrupt corresponds to an uncorrectable error at a logical CPU, the method further comprises: setting an uncorrectable error bit for each virtual CPU of the logical CPU; determining whether the error interrupt affects any of the virtual CPUs of the logical CPU; and when the error interrupt affects one or more of the virtual CPUs of the logical CPU, signaling the guest partition corresponding to the one or more of the virtual CPUs of the existence of the error interrupt to enable the guest partition to take corrective action.

13. The apparatus of claim 12 , in which the bus is a peripheral component interconnect express (PCI-e) bus.

14. The apparatus of claim 13 , further comprising a device coupled to the bus, in which the device generates the error interrupt.

15. The apparatus of claim 14 , in which the device is a storage device and the error interrupt corresponds to a failed data storage command.

16. The apparatus of claim 14 , further comprising a root port coupled between the bus and the device.

17. The apparatus of claim 16 , in which the processor is further configured to identify affected guests by determining owners of the device generating the error interrupt.

18. The apparatus of claim 12 , in which the processor is further configured to prevent reception of further error interrupts from the bus after receiving the error interrupt.